Does a common drug increase stomach cancer risk?

This story was in the papers this week, linking a very commonly used medication with a doubling in stomach cancer risk. A doubling in the risk of anything sounds bad, but what does it mean in reality?

“Acid reflux drug linked to more than doubled risk of stomach cancer”

The scientists, working in London, published a paper linking the use of proton pump inhibitors (PPIs) with cancer. As around 20 million people in the US take these drugs annually, usually to deal with heartburn, it is obviously very important to be aware of any potential harm they are doing. Indeed, several recent reports have emphasised that these drugs are not as safe as their maker’s say, although the levels of side effects is admittedly extremely low.This study found an increased risk of stomach cancer in people taking the drugs. Those on PPIs were 240% more likely to be diagnosed with the disease. However, to understand this properly, you have to realise the difference between relative risk (which this is) and absolute risk.For example, this increase in relative risk of 240% actually means an increase from 0.24% to 0.57% in the chances of an individual patient getting the disease. In other words, the likelihood of you being diagnosed with stomach cancer goes from a low risk to a slightly less low risk. So of the over 60,000 people included in this study, PPI use was associated with an extra 10 cases of stomach cancer. Most people would consider this acceptable for the benefit they get from the drug.While the increased risk for an individual is low, these drugs are among the most commonly prescribed in the world, as I mentioned above. So, for a population this big, a small increase in risk can result in thousands of additional cases of this disease.Extrapolations like this have to be taken with a pinch of salt however, as they are fraught with issues. For one thing, the study only followed people for 9 years, so it’s difficult to say much about a population of people taking the drugs. Most importantly however, these extrapolations make mass generalisations. Not all patients will be long-term users, while some will be on the drug for longer than the 9 years of the study. It is pretty clear however, that PPI use is associated with many additional cases of stomach cancer.It has to be pointed out though that this is a correlation, and we cannot say that the PPIs are causing the increase. I’ve discussed this in the past here, so I won’t go into detail on this.One last thing to mention is that although PPI use has increased dramatically since they were introduced in 1988, the number of stomach cancer cases has decreased by over 25% in the same time, due to other preventative measures we have put in place. This downwards trend is still continuing, so we can expect further drops in the coming years. Absolute risk v relative riskI showed above that even a large increase in risk of stomach cancer doesn’t mean many extra cases, and I have previously discussed the difference between absolute and relative risk in this blog post.Have a look at the diagram below. In both situations you have a 100% increase in relative risk. However, in one case this means your absolute risk goes from 1% to 2%. In the other it goes from 35% to 70%. Understanding this difference lets you be a lot more critical when reading numbers in the media!

Hot drinks and cancer

You may have seen a frankly terrifying headline this week:

“Hot drinks probably cause cancer, warns World Health Organisation”Telegraph, 15th June 2016

Almost every news source carried this story, and the headlines were universally similar to the one above. This story comes from a report by the WHO, which looked at the association between coffee and mate (a South American herbal tea) and various forms of cancer. In short, they found that there was no association between coffee or mate and cancer, but that the temperature of the beverage may be linked to oesophageal cancer. This, of course, is nothing to worry about. The report classifies hot drinks as “probably carcinogenic to humans”, group 2A in their classification. Other items in this category are the act of frying food, working as a hairdresser or barber, red meat, and working night shifts. This categorisation tells us about the hazard of hot drinks, but not about the risk.The words "hazard" and "risk" are regularly used interchangable, so the distinction between them is one that is lost on most people. Hazard is whether something can happen or not. Risk is the likelihood that it will happen. There is a hazard of crashing when you are driving with your eyes open or with your eyes closed. However, the risk is quite different in each of these cases. This report tells us about the cancer hazard of hot drinks, but nothing about the cancer risk, so the fact that hot drinks are on this list isn’t very informative.So what do we know about the risk of hot drinks. Firstly, this applies to drinks consumed at 65°C or above. So if you put milk in your tea or coffee, then you’re ok. Even if you don’t, oesophageal cancer isn’t very common, so even a big increase in risk wouldn’t translate into many more cases (see below for an explanation of this). If you want to do something to decrease your already small chance of getting oesophageal cancer, then consider stopping smoking, stopping drinking, eating more fruit and veg or losing excess bodyweight, all of which are known risk factors.The system of classification used by the WHO is unfortunately ripe for misinterpretation. It is almost impossible to prove a negative, so proving something definitely doesn’t cause cancer is difficult. The WHO has now classified 1,051 different things for their likelyhood of causing cancer. Of those, they have only rated a single one as “probably doesn’t cause cancer”. (For those who are interested, that one thing is Caprolactam, a compound used in the production of nylon).Nearly half of the rest fall into the “not classifiable” category because we just don’t have enough evidence to say either way. This may be because the evidence is inconclusive, or because studies have never been done. In reality, if the WHO analysed whether swivel chairs caused cancer, they would fall into this category rather than the "probably don't cause cancer"one, because we have never needed to study it.So by the WHO system, we can't say that the following don't cause cancer: chlorinated drinking water, caffeine, mobile phones, fluorescent lighting, hair colouring products, magnets or tea. They are in the “not classifiable” category. However, aloe vera, pickled vegetables and dry cleaning are all classified as “possibly cause cancer”. As you can see, this classification causes more confusion in the general public than anything else.When you understand the difference between hazard and risk it becomes far easier to interpret the constant health scare stories in the media. “Mobile phones may cause cancer” is a terrifying headline, until it is put into this context. The increased understanding of risk is a vital tool in the rational toolbox. And because large parts of the media don’t seem to possess this, it is one that we can get a lot of use out of! Absolute risk v relative riskI mentioned above that even a large increase in risk of oesophageal cancer doesn’t mean many extra cases. To understand this you have to understand the difference between relative risk and absolute risk.Have a look at the diagram below. In both situations you have a 100% increase in relative risk. However, in one case this means your absolute risk goes from 1% to 2%. In the other  it goes from 35% to 70%.To to bring it back to hot drinks, imagine a crazy situation where they give us a 50% increase in risk for oesophageal cancer. (Just to be clear, there is not a 50% increased risk with hot drinks, I made that number up as an example of a large increase.)The rate of oesophageal cancer is around 15 cases per 100,000 people, so your risk of getting it is 0.015%. A 50% increase in risk means that the rate would rise from 15 to 22.5 cases per 100,000. In this case your risk has now gone from 0.015% to 0.0225%, an increase of 0.0075%.You can see how an enormous increase in your relative risk (50%) can mean only a tiny increase in your absolute risk (0.0075%). So when you hear someone say that x increases your risk of cancer, your first question should be “but what is the risk of me getting that cancer in the first place?”. Once you know that you will have a far better idea whether the rest of the claim needs to be listened to.